DEVELOPMENT OF EQUATIONS FOR ESTIMATING YIELD LOSSES CAUSED BY MULTISPECIES WEED COMMUNITIES DOMINATED BY GREEN FOXTAIL [SETARIA-VIRIDIS (L) BEAUV]

Multiple regression equations were developed to describe the relations hip between percentage hard red spring wheat (Triticum aestivum L.) lo ss and the density and shoot dry weight of multi-species weed communit ies dominated by green foxtail [ Setaria viridis (L.) Beauv.]. Data we re collected over a 10-yr period from fields sown by farmers near Regi na, Saskatchewan. Weed densities averaged 470 plants m-2, with green f oxtail constituting of 85% of the total number of plants and 57% by dr y weight. Other important species included in the equations were stink weed (Thlaspi arvense L.), common lamb's-quarters (Chenopodium album L .), wild buckwheat (Polygonum convolvulus L.), and wild mustard (Sinap is arvensis L.). Including crop density as a variable made a significa nt improvement in the efficiency of the equations. Precipitation and g rowing degree-days (base 5-degrees-C) were related to the residuals fr om the analysis relating wheat loss to weed abundance. These environme ntal variables were significantly related to crop loss, but only for t he residuals of the equation where weed densities were used as indepen dent variables. Hyperbolic and sigmoidal equations were less efficient at describing the data than were multiple linear regression equations .